Download Lexical and Viewpoint Aspect in Kubeo

Lexical and Viewpoint Aspect in Kubeo*
Thiago Costa Chacon
Center for American Indian Languages – University of Utah
[email protected]
Key words: Aspect, Lexical Aspect, Tukanoan family, Kubeo, South America
1. Introduction
In this paper I explore how lexical aspect (stative vs. non-stative), understood as a
semantic and grammatical property of verb roots and stems in Kubeo 1, correlates with viewpoint
aspect (imperfective vs. perfective).
I follow Smith (1997) in using the term viewpoint aspect, instead of the traditional term
grammatical aspect, because of two reasons: (i) In Kubeo, there is no morpheme or syntactic
construction that can be analyzed exclusively as a perfective or imperfective marker. Hence, they
are not a grammatical category in the way described by Dahl (1985). They are rather a conceptual
category that can be inferred from the meaning of predicates. (ii) More recently in the theoretical
literature, perfectivity and imperfectivity have been used as semantic notions to describe
properties of predicates in languages with very different grammatical facts from Russian or Latin
– where perfective and imperfective have been traditionally used. Therefore, the term viewpoint
aspect achieves two goals in this paper: it captures an important fact internal to Kubeo grammar
and make the analysis of this language available for cross-linguistic comparison.
The specific way that lexical and viewpoint aspect correlate in the language is stated in
(1) below:
(1) A predicate that is headed by a stative stem has an imperfective grammatical aspect;
while a predicate headed by an eventive stem has a perfective grammatical aspect.
An introduction to the Kubeo data to support the analysis above will be given in section
1.1 below. In section 2, when I deal more closely with the typology and theory of aspect, the
parallelism between lexical aspect of verbs and viewpoint aspect of predicates as whole will be
better motivated. In section 3 and 4 I return in more detail to the Kubeo facts. Section 5 is the
conclusion of this paper.
Throughout this paper I present arguments against previous analyses of Kubeo. More
specifically, in my description of the Kubeo facts below I argue against the analysis of Morse &
Maxwell (1999) and Hollinger et al. (2000). They have argued for a split tense system, where
stative verb stems when combined with tense-evidentialiy-mood-person 2 morphemes (henceforth
* I would like to thank Lyle Campbell and Aniko Csirmaz for their comments in previous versions of this paper.
1
Kubeo is an Eastern Tukanoan language (Mason 1950), spoken by 4,000 people on the Vaupés and Ayari rivers, at
the Brazilian and Colombian border area. Data for this study was collected with the southern most Kubeo dialects
during two fieldwork seasons (Summer 2008 and 2009). This study is part of an ongoing investigation towards the
completion of a reference grammar of Kubeo, which is also my PhD dissertation topic.
2
Kubeo verbal inflectional morphology has a high degree of fusion, where a single morpheme refers to more than one
inflectional category. Inflectional categories in Kubeo verbs include Tense, Evidentiality, Mood and Person. I will use
the acronym TEMP to refer to the collection of inflection categories in Kubeo verbs. There are two contrastive
inflectional paradigms in Kubeo verbs based on the tense system: non-past and past. Since tense is always coded in
combination with other inflectional categories, I will use the terms ‘non-past TEMP’ and ‘past TEMP’ to refer
specifically to each set of forms, or just TEMP to refer to the verbal inflectional morphemes in general.
TEMP) yield present tense, and eventive verb stems plus TEMP morphemes yield past tense.
I argue for a split aspectual system, which seems to be a more natural explanation for the
Kubeo facts, instead of a split tense analysis, for theoretical, typological and Kubeo-internal
grammatical reasons that will become clear in the course of this paper.
1.1 Introduction to Kubeo data
The categories of tense and aspect are coded within the VP in Kubeo. There are,
however, many points that make Kubeo tense and aspect system typologically uncommon. A
brief introduction to how these categories correlate in Kubeo is necessary before we proceed to a
more detailed account.
Verb roots in Kubeo can be divided into two classes, depending on the lexical aspectual
properties of their meaning: stative and eventive (non-stative) classes3 . When combined with
specific TEMP morphemes (stative root + TEMP, or eventive root + TEMP) there is a clear
distinction in meaning at the predicate aspectual level between the two constructions. Consider
(2) below, where verbs are combined with non-past TEMP4:
(2)
a.
b.
da-bi5
he come-N.PST.3.MSC
‘he has arrived’
eventive verb root
!! toro-h!-bi
he clourful-VBLZ-N.PST.3.MSC
‘he is happy’
stative verb root
As it can be seen, the combination of an eventive root plus non-past TEMP morphemes in
(2a) yields a predicate with perfective meaning and a situation time reference in the past –
although tense is morphologically unmarked, therefore non-past (also, in section 4.6 I show how
perfective predicates with non-past TEMP can also code situations with future time reference).
On the other hand, the combination in (2b) of a stative verb root plus the same TEMP morphemes
as in (2a) yields a predicate with imperfective meaning and a situation time reference in the
present.
For an eventive verb root to code a situation time reference in the present, it must be
combined with some derivational device, which derives a stative stem6 . In (3) below, the
derivational morpheme –i ‘stative’ assumes this function, yielding a non-progressive stative
predicate7 :
(3)
3
da-i-bi
come-ST -N.PST.3.MSC
Verb class can also be defined also by grammatical criteria (such as morphological selection, cf. section 3 below).
A note on the orthography of this paper: <ç> = /t"/; <y> and <ñ>/j/; nasal syllables are marked by a tilde over the
vowel or by a nasal consonant <m, n, ñ>. Stress is marked on unpredictable syllables, otherwise it falls in regular
iambic intervals, parsed from the left to the right. Primary stress falls in the left most stressed syllable.
5
1 ‘1st person’; 2 ‘2nd person’, 3 ‘3rd person’, AN ‘animate class’, ANPH ‘anaphoric’, ANT ‘anteriority nominalizer’, C
‘inanimate count’, CAUS ‘causative’, C ‘classifier’, CLT ‘clitic’, CONV ‘converb’, COP ‘copula’, DIM ‘diminutive’, DUR
‘durative’, DYN ‘dynamic’, EXCL ‘exclusive’, FEM ‘feminine’, FOC ‘focus’, HAB ‘habitual’, IF ‘if/conditional’, IMPR
‘imperative’, IN ‘inanimate class’, INCL ‘inclusive’, INTRJ ‘interjection’, IRR ‘irrealis’, LOC ‘locative’, MS‘inanimate
mass’, MSC ‘masculine’, MY ‘1st singular possessive’, N ‘non-’, NEG ‘negative’, NMZ ‘nominalizer’, OBL ‘oblique case’,
ORG ‘origin’, PASS ‘passive’, PERF ‘perfective’, PERM ‘permissive’, PL ‘plural’, POSS ‘possessive’, POST ‘posteriority
nominaizer’, PSNT ‘present’, PST ‘past’, Q ‘question marker’, RSLT ‘resultative’, S ‘singular’, SIM ‘simultaneous’, ST
‘stative’.
6
See section 3 for the distinction of roots and stems.
7
For a better treatment of the function and meaning of the morpheme –i ‘stative’, see section 3 and 4 below.
4
‘he comes’ (lit. ‘he is on his coming’)
On the other hand, for a stative root to code a perfective situation with past time
reference, some derivational process must also occur, making an eventive stem. In (4) below, the
derivation morpheme –te ‘dynamic’ assumes this function, yielding a perfective, inchoative,
punctual event:
(4)
kari-de
curren.time-OBL
‘he got happy today’
!! mea-rõ
he good-IN.SG
toro-h!-te-bi
clourful-VBLZ-DYN-N.PST.3.MSC
Also, in connection to –i ‘stative’ and –te ‘dynamic’, there are other derivational
morphemes that exhibit a similar function of deriving a verb root to an eventive stem or a stative
stem. These other morphemes can be analyzed as a matter of their intrinsic grammatical function
(valency, modality, etc), the resultant lexical aspect of a stem (stative vs. non-stative) and the
default viewpoint aspect of the predicate (perfective vs. imperfective).
The Kubeo system grows more complex as we start describing the correlation between
lexical and viewpoint aspect with past TEMP morphemes. Eventive verb stems combined with
past TEMP yield situations with past time reference, while stative stems combined with past
TEMP yield a predicate with generic meaning. Consider (5a) and (5b). A more detailed analysis
of non-past and past tense will be given in section 4.
(5)
a.
b.
yawi makarõ-i
k!-ame
jaguar jungle-LOC exist-PST.3.MSC
‘the jaguar lives in the jungle’ (as a universal statement)
STATIVE STEM
yawi makarõ-i
k!-te-ame
jaguar jungle-LOC exist-DYN-PST.3.MSC
‘the jaguar was in the jungle’ (when the speaker saw it)
EVENTIVE STEM
Generic sentences such as (5a) bear marking which distinguish them from habituals in
Kubeo, an issue that will be discussed in more detail in section 4.
A summary of the facts discussed in this introduction is given in the table below. Further
details, commentaries and discussion of minor exceptions will be given in section 4:
Table 1. Correlation between TEMP, lexical aspect, viewpoint aspect and situation time
reference
TEMP Morpheme Set Lexical Aspect of the Viewpoint Aspect
Situation time
Stem
reference
Non-past
Stative
Imperfective
Present
Eventive
Perfective
Past or Future
(non-present)
Past
Stative
Imperfective
Generic
reference
Eventive
Perfective
Past
2. Tense and aspect
For the next subsections, I will present how typological literature has treated viewpoint
(grammatical) and lexical aspect8. In discussing lexical aspect, I will focus primarily in the
distinction of states and non-states, due to its relevance for the Kubeo tense and aspect system.
Later, in 2.4 I will try to motivate a semantic natural link between the notions of stativity –
imperfectivity, and eventivity - perfectivity, in order to support the correlation between lexical
and viewpoint aspect in Kubeo theoretically.
2.1 On the relation of TIME, tense and aspect
Tense and aspect are two grammatical categories that function to express the semantic
notion of TIME in languages (cf. Comrie 1976; Dahl 1985; Binnick 1991; Klein 1994). Comrie’s
(1976:5) semantic distinction between aspect and tense uses two parameters for presenting a
situation in time: aspect refers to the “situation internal time”, i.e. how a situation is described in
relation to its temporal internal constituency; tense refers to “situation external time”, i.e. how a
situation is presented in relation to an external point of view, such as the speech time. According
to Comrie’s definition, tense systems are distinct from aspect systems by being primarily a deictic
category.
There is an important distinction between tense and time reference in Comrie’s 1985
proposal. While tense is a grammatical category, time reference is a broader semantic operation to
locate any linguistic expression in time. So temporal adverbs, such as today or tomorrow, are
devices of time reference, but not of tense. The fact that time reference and tense are distinct
entities can be easily demonstrated by the English present tense, which may have both present
(6a), (6b) and (6c), and future time reference (6d), or even performative meaning (6e):
(6)
a.
b.
c.
d.
e.
I see your point
I am now writing you a letter.
I am writing letters on Tuesdays.
I am writing a new letter tomorrow.
I promise to write you.
This correlates with the distinction between situation time and the linguistic statements
made by speakers about a situation (Klein 1994). While situations are the real-world facts, the
linguistic statements are representations of real-world facts. Suppose an event that took place in
the past can be ‘neutrally’ conceptualized as MONKEY EAT BANANAS. This event can be either
described as a progressive event The monkey was eating bananas, or as a perfective event The
monkey ate bananas. Therefore, there is in some degree an arbitrary link between the real world
event time and the event description made by speakers (cf. Binnick 1991:184 for a more profound
way to motivate this distinction). It seems that the classical Sassurean concept of linguistic sign
stems from a similar observation.
Aspect simply refers to the description of the event itself, whether it is completed or not,
whether it is punctual or continuous, etc.
(7)
a.
b.
When you arrived, I was leaving home.
When you arrived, I left home.
In (7) the main clauses describe the event of leaving in respect to the punctual event of
arriving. Since arriving is a single point in the discourse time, the main clauses can either overlap
or not overlap (precede/follow) it temporally. In (7a) leaving was taking place during the
moment when there was an arriving event, hence overlapping with the arriving event. In (7b) the
basic interpretation is that there was an event of leaving after the event of arriving, hence not
8
Unfortunately, due to space limitations, I will not be able to show full examples of all the languages surveyed in the
literature that are important for illustrating and supporting my analysis. This will be done in a future study, though.
overlapping. This is due to the fact that the progressive form in (7a) has enough ‘temporal
extension’ to allow overlapping, while this is not the case with the perfective form of (7b).
Dahl’s 1985 study showed that distinctions such as perfective and imperfective are
typologically universal. A fair number of tenseless languages has been described, where TIME
information is coded by aspectual rather than by tense (Comrie 1985). Olsen (1997:5) notices
‘few languages lack aspect distinction, while many languages lack tense’. These facts can be
indirectly related with the fact that, diachronically, aspect systems seem to feed tense systems (cf.
Comrie 1985, Klein 1994).
2.2 Aspectual categories
With the general background from previous section in mind, I turn now to address
relevant issues about aspect and tense that will give us the basis for an analysis of Kubeo.
According to Binnick (1991), as a conceptual category, aspect was a familiar topic to
classical grammarians, as for example in Varro’s distinction between Latin perfectum ‘complete
action’ and imperfectum ‘incomplete action’. Later, Slavicist adopted an analogous distinction for
Slavic languages (Binnick 1991:136).
In discussions of aspectual categories, perfective and imperfective are taken to be the
most basic categories, both typologically and semantically. Comrie (1976:16) defines perfective
and imperfective aspects as two bipolar ways of describing situations: perfective describes a
situation as a complete event; imperfective describes a situation with internal temporal structure.
Smith (1997:65) talks about families of aspectual ‘viewpoints’. Perfective viewpoints are defined
by presenting a situation closed informationally, while imperfective viewpoints present situation
open informationally. Dahl 1985 defines perfective aspect as describing situations as “a single
event, seen as an analyzed whole” (1985:78), though he does not give any positive definition of
imperfective aspect.
With respect to the notion of perfectivity, Comrie 1976 emphasizes the difference
between “describing a situation as a complete event” from “a completed event”. Contrary to
“completed”, a complete event does not necessarily imply that the event culminated or was
successfully completed. Comrie (1976:18) states that in Russian and Ancient Greek there are
situations with future time reference that are expressed by the perfective/aorist.
In other cases, the issue of completion (telicity) or non-completion of predicates with the
perfective aspect seems to be due to lexical aspect (or to situation type as defined by Smith
(1997:68)). Telic situations, such as accomplishments tend to exhibit completion, while atelic
situation tend to exhibit termination, but not completion. Nevertheless, every predicate with the
perfective aspect tends to exhibit “initial and final points” (Smith 1997:66).
Smith (1997:110) argues that there is a pragmatic principle that constraints situations with
present time reference to include the endpoints of the situation, which can explain Dahl’s (1985)
observation that the perfective aspect is absent in present time reference typologically. This can
also explains the Kubeo facts introduced in section 1.1 (cf. Table 1).
Closer to our interest in this paper is Smith’s (1997) interpretation of ‘non-canonical’ use
of the perfective aspect with stative verbs. She argues that in some languages an inchoative
interpretation is available for the combination of STATIVE VERB + PERFECTIVE, while in some
other languages this interpretation is not available, and the combination results in describing a
‘closed state’ (Smith 1997:69). French is mentioned by Smith as one kind of language in which
the combination of STATIVE VERB + PERFECTIVE codes a closed state, as in the example below:
(8)
a.
Marie a vecú à Paris
Marie lived in Paris
b.
# Marie a vecú à Paris et elle y vit encore
Marie lived in Paris and she still lives there.
Cuzco Quehcua (personal data) seems to be the type of language where the combination
of STATIVE VERB + PERFECTIVE entails a change of state. Welmers (1973:347-8) present evidence
that many Niger-Congo languages behave similar to Cuzco Quechua.
However, it turns out that the two interpretations – closed states and inchoatives - are
available within the same language as well, as we can see on the Portuguese examples below:
(9)
a.
Ela dormiu
a
noite inteira
she slept
the
night entire
‘She slept for the entire night’
b.
Silêncio! A
criança dormiu agorinha
Silence the
child
slept
now-DIM
‘Silence! The child fell asleep just now’
In Kubeo, there seems to exist the two possibilities as well. I show a stative predicate in
(10a) and its non-inchoative, closed state interpretation in (10b). In (10c), the same verb form
describes an inchoative situation, which is similar to the Portuguese example above. Compare
sentences (2b) and (4) (above) for another inchoative interpretation with a stative verb or see
section 4.1 for additional examples:
(10)
a.
y!-re h!h!-w!
I-OBL cold-N.PST.N.3.AN
‘I am cold’
b.
kari
yãmi-ne y!-re h!h!-te-w!
current.time night-OBL I-OBL cold-DYN-N.PST.N.3.AN
‘I felt cold last night’
c.
kari
yãmi
current.time night
‘this night got cold’
h!h!-te-w!
cold-DYN-N.PST.N.3.AN
Notice that the constituent ‘night’ in (10c) is the subject and it does not have the oblique
marker that was present in (10b) when it was not a core argument. In (10a) the argument ‘I’ has
an oblique marker, and it is not the subject. It is in fact some kind of dative argument, like the
Spanish dative constructions as in me gusta Juan ‘I like John’ (lit. ‘John pleases me’), where me
‘I’ is not the subject.
The Perfect is classified by Dahl 1985 as a kind of grammatical category that needs an
imprecise definition, as some point in between aspect and tense. In general, the Perfect is
characterized as describing situations that occurred prior to the reference time. Comrie (1976)
defines the perfect as a kind of aspect that presents ‘some past situation with present relevance’,
the latter being vaguely defined. This is related to notions that usually involve the Perfect, as
resultative, subject experiencer, etc. Moreover, Smith (1997:108) defines situations marked by
the Perfect as denoting a resultant state, although the aspect is perfective. For her analysis, the
Perfect would be similar to habituals marked by perfective aspect, i.e a derived state.
Dahl (1985:139) states that the perfect is usually marked by syntactic morphemes,
usually some type of copula or ‘have’ auxiliary in present tense plus a participle verbal form. This
composition corresponds to Comrie’s notion of ‘past situation’ (coded by the past participle) with
‘present relevance’ (coded by the tense in the auxiliary form).
Kubeo has some forms that may resemble the Perfect category cross-linguistically, both
in form and meaning (cf. section 4.1)
Imperfective and its correlates code predicates with no delimited temporal boundary,
thus the information they code is open for interpretation with respect to both or either of its
terminal and starting points (cf. Smith 1997:73).
The most common types of imperfective aspects are the progressive and the imperfective
proper. Sentences in the non-progressive present tense are usually analyzed as imperfective. In
Russian, the imperfective aspect is the only one available for the description of present situations
(cf. Comrie 1976; Smith 1997). Contrasts of imperfective and perfective aspects mostly occur in
past tense (what could be due to the pragmatic constraint in relation to the use of perfective aspect
in present time reference). Smith (1997:74) states that progressives have “meanings that do not
arise for other types of imperfective. Nuances of activity, dynamism […] are often associated
with the [progressive]”. This could be the reason why progressives and states are usually in
conflict (though Mufwene 1984 has a different opinion).
Comrie (1976:26) supports his analysis of habitual as a subcategory of imperfectivity
from evidence in several languages that have a single imperfective marker to code both
continuous and habitual events. Smith (1997:50-51) argues that habituals are derived states. They
have semantic properties of states, but sometimes can have in addition syntactic properties of
events, for instance Susan rode her bike during the whole summer, where perfective aspect is
used to describe a habitual sentence. Comrie 1976 also gives similar examples. However, this
would argue against the analysis of habituals as a subcategory of imperfective.
There is evidence that habitual might not be an aspectual category at all (cf. Flip &
Carlson 1997). Comrie (1985:51) reports that in Burmese and Dyirbal, both tenseless languages,
habituals are distinctly coded by realis and irrealis mood, respectively. The connection between
irrealis modality and imperfective categories is well discussed in the literature (cf. Comrie 1985;
Fleischman 1995; Smith 1997). Kubeo seems to support this too (cf. section 4.1).
From several languages, there is evidence that habituals, generic and imperfective belong
to the same category. Filip & Carlson (1997:17) states that ‘generics are aspectually stative […]
and the aspectual character of imperfectives seems to be more semantically compatible with
stativity than that of perfectives’. They also claim that generics are a category of its own (1997:1).
Accordingly, Smith (1997:33) analyzes generic predicates as one kind of derived states.
From the Kubeo perspective, habituals and generics are different grammatically and
semantically. The latter is a kind of stative predicate, while the former is grammatically perfective
(cf. section 4).
2.3 Lexical Aspect
Early insights of lexical aspect can be traced back to the Aristotelian classification of
situations (Binnick 1991:170). Binnick (1991, chapter 6) shows that lexical aspect belongs to a
totally different tradition of studies from Aktionsarten, and both are also conceptually different,
despite the fact that some scholars use the terms as synonyms.
Lexical aspect, on the other hand, relates to the much-debated classification of
eventualities, such as in Vendler’s (1957) seminal classification of situations in states,
achievements, accomplishments and activities. Vendler’s (1957) basic idea is that ‘the fact that
verbs have tenses indicates that considerations involving the concept of time are relevant to their
use. […] [T]here is another, a more subtle dependence on that concept: the use of verb may also
suggest the particular way in which that verb presupposes and involves the notion of time’
(Vendler, 1957:143). By what he calls verb schemata, verbs can be defined in 4 classes
depending on how the concept of time is relevant to particular verbs, as he states:
Table 2: Vendlerian classification of lexical aspect
Lexical Aspect
Time Definition
Examples
STATES
ACTIVITIES
nonunique and indefinite time
instants
not unique or definite time
periods
ACHIVEMENTES
unique and definite time instants
ACCOMPLISHMENTS
unique and definite time periods
Peter loves Mary.
She is sick.
I pushed the cart for two
hours.
He ran in the park.
John reached the top of the hill.
He arrived late.
The kid ate a sandwich.
Kubeo seems to give grammatical relevance to two types of situations: state and events.
Events are a superordinate class that covers all non-stative situations: activities, accomplishments
and achievements. Sometimes the class of events goes with another name, such as Mourelatos’
(1978) occurrences; dynamic as in Comrie (1976); or more commonly and inexactly as acitivy.
However, usually event is used as a ‘negative category’ - that of non-states.
On the other hand, it seems that every language has stative and non-stative as the most
important lexical aspectual classes in its grammar. For instance, Western European languages
also tend to oppose states and non-states, such as the contrast in the interpretation of the sentences
below with present tense:
(11)
a.
Susan owns a farm.
STATE
PRESENT AND ONGOING SITUATION
b.
Peter rides horses.
ACTIVITY
HABITUAL
c.
My wife arrives late.
ACHIEVEMENT
HABITUAL
d.
I build houses.
ACCOMPLISHMENT HABITUAL
The data shows that all classes of non-states behave similarly with respect to the predicate
interpretation, as opposed to the stative predicate. This also seems to be the case for many other
languages. For instance, Cuzco Quechua (personal data) also distinguishes states from non-states,
but in a slightly different manner: in non-past tense, non-stative predicates unmarked for aspect
can present either a habitual interpretation, past or future time reference. Welmers (1973:346)
gives evidence that most Niger-Congo languages distinguish two types of verbs, states and
events, where eventive predicates in non-past tense have usually a Perfect interpretation, with the
past time reference. This also seems to be the case with a number of Creole languages (cf.
Mufwene 1984).
Smith (1997) and Binnick (1991) state that the key feature that distinguishes states from
events is that the latter have intrinsic temporal boundaries. Even activities or other durative events
have temporal boundaries, as Smith (1997:35) states “the initial endpoint is a change from a state
of rest; the final endpoint is change into a state of rest, even if the event is telic or atelic”.
The property of temporal boundaries is also the key factor that can apply to what Smith
calls derived states, such as progressive, habituals, and others. Crucially, the progressive is
understood as a derived state, though it has a dynamic meaning, hence its notion as one type of
derived state (or “grammatical stativity” as Mufwene 1984 calls it) is mostly due to the fact that
the progressive does not give explicit information about its onset or terminus.
2.4 The correlation of lexical and viewpoint aspect
Mourelatos 1978 states that traditional studies of lexical aspect “had [not] realized that the
distinctions they sought to articulate had long been studied by linguists under the heading of ‘verb
aspect9” (Mourelatos 1978:418).
He emphasizes that lexical aspectual distinctions in terms of the Vendlerian classification
“will be misdescribed if it is thought that they arise mainly from the semantics of individual
9
Where Mourelato’s term ‘verb aspect’ corresponds to grammatical or viewpoint aspect.
verbs, when in fact they involve fundamental linguistic categories reflected partly at the lexical
level and partly […] at the morphological and syntactic level” (1978:419) (my emphasis). He
mentions Vendlerian stative verbs such as know (1978:419) and see (1978:422), which seem to
act strangely in their aspectual behavior in past tense. Compare the stative use in the sentence I
know him and its achievement use in Suddenly, I knew the answer. What Mourelatos suggests is
that there is no change in the verb itself, but a change in the tense and aspect from one sentence to
the other (from imperfective to perfective), which is responsible for the ‘change’ in the lexical
aspectual category of the verb.
Mourelatos (1978:429) argues for an interesting correlation of perfective aspect with
eventive predication and count nouns, and imperfective aspect with stative predication and mass
nouns.
Disregarding the generalization over the noun domain, the correlation between
stative/eventive predication and imperfective/perfective aspect seem to be very suitable for the
analysis of Kubeo (cf. section 1.1). Interestingly, the pair of examples (2b) and (4) is analogous to
the know/knew distinction drawn by Mourelatos. This is mostly because some stative verbs in
perfective aspect can code inchoative predicates (cf. Smith 1997:69).
In section 2.2 and 2.3 I showed cases in the literature that point to a correlation of stativity
with imperfective, progressive, habitual and generic predicates (also called by Smith 1997 as as
derived states).
All of these considerations seem to show that the parallelism between lexical and
grammatical aspect in Kubeo is well founded in typological and theoretical literature. The
explanation for these facts can be stated semantically as:
(12) a.
Events are definite and discrete situations. Perfective aspect describes
situations with definite and discrete temporal boundary in the discourse timeline.
b.
States are indefinite and non-discrete situations. Imperfective aspect describes
situations with indefinite and non-discrete temporal boundaries in the discourse
timeline.
While this is both a theoretical generalization and an analysis proposed for Kubeo, this
interpretation can also be supported from facts in unrelated languages, such as the languages
mentioned in previous sections and also in Igbo. Emananjo (1991) states that in Igbo there is a
past vs. non-past tense distinction, and in verbs with non-past and “neutral aspect” the perfect is
the default interpretation for non-stative verbs, while stative verbs have a timeless interpretation
(1991:135). Stative verbs, such as ‘to be big’, ‘to be beautiful’, when combined with completive
(perfective) suffix, describe a change of state or inchoative situation. In sum, the lexical aspect of
the head of the predicate, whether a stative or eventive verb, implies the grammatical aspect,
defining it either as imperfective or perfective, respectively. Emananjo’s (1991) analysis of Igbo
is very similar to the one proposed for Kubeo, and neatly matches the generalizations in (12).
2.4.1 Tense and correlation of lexical and grammatical aspects
Tense alone cannot be understood to be a factor for giving a definite time schema for an
eventuality. For instance, Hatav 1989 says that stative verbs may or may not be part of a narrative
sequence, depending on how the temporal limit of the situation is coded. She argues that in order
for a state to be part of a narrative sequence, the stative verb must be delimited either by
perfective aspect or some time adverbial, such as for X hours (Hatav 1989:497-8).
Consider the contrasting sentences below:
(13) a.
He lived in China for ten years and moved to Japan
b.
#He lived in China for ten years when I met him
c.
He was living in China when I met him.
All of the sentences above are in the past tense. Sentence (13a) has both clauses in the
perfective aspect, and hence both are understood as a sequence of events. Sentence (13b) is not
accepted when the main clause cannot code a time interval, at which a second event took place. If
the conjunction were before or after and did not allow overlapping, the sentence would be
felicitous. On the other hand, (13c) has the main clause in the progressive aspect, with indefinite
temporal boundaries, which allows for a second event to take place within its (loose) time
interval.
This shows that tense alone is not crucial to make eventualities more or less definite and
discrete. This is related to the ontological properties of tense, more crucially to the fact that “tense
describes the situation time from an external point of view” (Comrie, 1976), i.e. by always
describing situations in respect to a deictic center that is external to the situation itself, tense does
not contribute to the notion of definiteness.
This is another argument for not ascribing the Kubeo facts described in 1.1 to a split tense
system, as it was mistakenly argued for by Morse & Maxwell (1999) and Hollinger et al. (2000).
In fact the Kubeo tense is very simple and similar to the majority of languages with a past/nonpast distinction once one realizes that the most important elements in this system are not related
to tense, but to aspectual relations and time reference constraints. One difference between Kubeo
non-past tense and Western European languages non-past tense category is that in Kubeo this
tense can have past time reference, as explained in 1.1. There are additional idiosyncrasies in
Kubeo past tense that can contribute to a general theory of tense systems, though – of course –
they are not mysterious in any ways.
3. Lexical Aspect and Grammatical Facts in Kubeo
In this section, I discuss in 3.1 how a three level distinction in Kubeo grammar can be
motivated based on the semantic and grammatical facts of stative and eventive situations: the
lexical, the grammatical and the predicate levels.
3.1 Morphological selection
Stative and eventive verbs in Kubeo fall into semantic and grammatical classes that
encompass all verbs in the language. A basic grammatical distinction between stative and
eventive verbs is seen in their behavior with respect to nominalizations, which in Kubeo involves
three types of time reference in general: anteriority, simultaneity and posteriority. (consider
Table 3 below)10.
The distribution of each set of forms and verb roots is asymmetrical. For instance, stative
verbs do not get combined with nominalizers of simultaneity. To refer to a present sate of a
referent, stative verbs must get combined with anteriority nominalizers (14), while eventive verbs
take the simultaneity set (15). For the posteriority set there is no restriction on verb root class.
Table 3 Nominalizers based on temporal reference, verb root class and gender categories
TEMPORAL
ANTERIORITY
SIMULTANEITY
POSTERIORITY
REFERENCE
VERB ROOT CLASS
MASCULINE
FEMININE
ANIMATE PLURAL
INANIMATE COUNT
10
STATIVE
EVENTIVE
EVENTIVE (only)
BOTH
-k!
-ko
-riw!
-rõ
-k!
-ko
-rã
-y!
-yo
-iw!
-ino
-k!y!
-koyo
-rãhiw!
-kino
Kubeo has a very complex set of deverbalizers that also have homophonous forms in noun and adjective inflectional
paradigms. The particular nominalizers in Table 3 are the ones used as complements of copulas in the language.
INANIMATE MASS
INAN. + CLASSIFIER
-e
-ri-cl.
-re
-iye
-i-cl.
-kiye
-ki-cl.
So, a stative verb takes a nominalizer of anteriority to refer to a present state of its
referent (14), while an eventive verb takes a nominalizer of simultaneity (15):
(14)
(15)
a.
mea-ko=be
nomi-o
good-ANT.FEM=3.AN.SG.COP woman-FEM
‘the woman is beautiful’
b.
bia=ka
h!me-ni=ka
chili-cl.SM.RND green-IN.NMZ =CL .SM.RND
‘a green chili’
a.
õ
kore-yo=be
she
wait-SIM.FEM=3.AN.SG.COP
‘she is waiting/taking care of’ or ‘she is the one who waits/take care of’
b.
w!-i=k#
fly-ST=CL.BIG.HOLLOW
‘an ariplane’
One can see that the forms of simultaneity in Table 3 are marked by –i ‘stative’, except
for the masculine and feminine singular forms, where –i ‘stative’ has grammaticalized, fused as a
glide with the gender forms. The fact that –i ‘stative’ is pervasive in the simultaneous nominalizer
forms and the fact that only eventive verbs get combined with this set of nominalizers are
intrinsically related to lexical aspect and time reference, i.e: eventive verbs need to get
combined with some stative element in order to refer to present, ongoing situations11.
A second factor from morphological selection that corroborate the grammatical
distinction of states vs. events is the fact that the morpheme –te ‘dynamic’ can only get combined
with a stative root, stative stem or a noun, but not with an eventive root or eventive stem. This a
diagnostic implicit in the analysis in section 4, when I will refer to stative or eventive stems based
in this criteria and others presented in section 3.1, 3.2 and 3.3.
3.2 A three level distinction: root, stem and predicate
In Eastern Tukanoan languages, bare verb roots and stems may coincide in their form.
While Kubeo verb roots are monomorphemic, verb stems can be monomorphemic or
plurimorphemic, depending whether there is or not a derivational morpheme (not a nominalizer)
before inflectional TEMP morphology. The morphological template of Kubeo verb is given in
(16):
(16)
[INCORPORATED NOUN[[[ROOT]STEM]]INFLECTIONAL MORPHEMES]=CLITICS
LEXICAL UNIT
DERIVATION
TEMP
The derivation morphemes adjacent to the root may alter its semantics, valency,
modality, etc., but not its grammatical category as verbs. Since they only derive verbs from other
verbs, they are a type of non-category changing derivational process. On the other hand, they can
make a stative stem from an eventive root, as well as an eventive stem from a stative root. This is
11
If negation, another stativizer morpheme, follows an eventive the stativer anteriority set of nominalzers is used.
illustrated in examples from (2) to (5) above for the function of –te ‘dynamic’ and –i ‘stative’.
Other verb class changing morphemes are given in the Table below:
Table 4: class changing morphemes
Grammatical Function
Derive eventive stems
Causative
X
Authoritative
X
Benefactive
X
Habituals
X
Irrealis
Negation
Derive stative stems
X
X
A negative suffix that gets combined with a stative root does not change its verb class,
neither does the causative change the verb class of an eventive root. In the absence of the
morphemes in Table 4, or –i ‘stative’ and –te ‘dynamic’, the lexical aspect of the stem is the same
as the lexical aspect of the root, as it was illustrated in section 1.1.
In this regard, one must distinguish two levels of semantic and grammatical classes of
verbs in Kubeo: one is the root level, where roots are divided into two lexical classes, stative and
eventive roots. The other is the stem level, where stems are divided into two grammatical
classes, stative and eventive stems.
The reason for calling the root level distinction lexical class is because this division is
ultimately important to the type of morphology that depends on lexical selection, such as
nominalization, word formation, etc. On the other hand, grammatical class is the distinction on
the stem level because it has implication into deeper grammatical levels, such as grammatical
aspect, valency, phrase heads, syntax, etc.12
A third level is the predicate level, which is ultimately semantic, since this is the
propositional level, but it relies on the grammatical facts from the stem plus TEMP morphemes. It
is at the predicate level that one can refer to imperfective and perfective aspect.
3.3 Some on the syntax of stative and eventive verbs
A stative verb functioning as an auxiliary verb requires the main verb to be marked by –i
‘stative’. An eventive verb functioning as an auxiliary verb requires the main verb to be marked
by the converb suffix –ri!.
For instance, the ‘frustrative’ construction is formed by the auxiliary eventive verb dú
(which when occurring independently has the meaning of ‘to release’, ‘to escape’), while the
main verb is formed by the verb root plus a converb suffix, as in (17):
(17)
k#-r$
dú-ame
y!-re
bite-CONV FRUST-PST.3.MSC I-OBL
‘he tried to bite me (but he did not)’
On the other hand, the ‘desiderative’ construction is formed by the stative verb ! ‘to
want’, functioning as an auxiliary, while the main verb is followed by –i ‘stative’:
(18)
12
karí-de
hãrãw! hoa-rõ
!%yei ã-i
!-te-w!
current.time-OBL day
long-IN.C grape eat-ST want-dyn-N.PST.3.AN
‘I wanted to eat grape for the whole day’
I use the term ‘grammatical’ closer to the classical use of this term.
We thus have evidence that the stative vs. eventive distinction is pervasive in Kubeo
grammar, from the semantics of verb roots, to the morphology, syntax and the meaning of
predicates.13
4. Kubeo tense and aspect system: a deeper look
In this section I focus on Kubeo predicates. The discussion will be based in the following
topics: in 4.1 I discuss the contrast of states, generic sentences and habitual predicates; in 4.2 I
show how the progressive contrasts with states of ongoing events marked by –i ‘stative’; in 4.3
resultatives (constructions that resemble the Perfect cross-linguistically) are discussed in contrast
to perfective events; in 4.4 I give a more detailed account of the morpheme –te ‘dynamic’; section
4.5 discuss the connection of negation, states and inchoative predicates; section 4.6 describe
issues related to temporal reference, aspect and tense14.
4.1 States, Generic and Habitual
Kubeo grammar has distinct constructions for stative, generic and habitual predicates. A
basic distinction between stative and generic predicates is that the former is marked by non-past
TEMP morphemes, while the latter is marked by past TEMP.
A stative stem using non-past TEMP can code three types of states: stable states,
transitory states and states of ongoing events (the former will be discussed in section 4.2).
Situations that code transitory states have a stative root as the head of a stative stem. Consider
the examples below:
(19)
a.
ihi-Ø-di
kari-de
pain-N.PST-Q
curr.time-OBL
‘does it hurts? Now it hurts’
ihi-w!
pain-N.PST-N.PST.3.IN
b.
ape-no
!-k!-r!! bí
mea-w!!
çiã!
other-IN.C want-ANT .MSC good-N.PST.1SG msc.vocative
‘do you want some more (another one)? No I am fine, man’
The set of forms of the non-past tense above requires witnessed or sensorial experience of
the speaker; therefore, it is suitable for the description of present, momentaneous situation. This
interacts with the semantics of stative roots, reducing their lexical time to hold transitorily or
momentaneously.
Stems headed by a stative root can also appear in predicates of more stable states. This is
a shift in the more common interpretation. The sentences in (20) below present the verb ihi ‘to be
in pain’ with non-transitory, stable state meanings. They contrast with the transitory meaning in
(19a). In (20a) the stem is headed by a stative root; in (20b), the stem is headed by –i ‘stative’
which follows the present habitual suffix:
(20)
a.
õ ã-yo-re
kai=n!m!-a
yapi-b!
iji-v!
she eat-SIM.FEM-OBL every=day-PL stomach-CL .TUBE pain-N.PST.3.AN
‘when she eats, her stomach hurts’
b.
bia
õ-i
ã-ru #med!
iji-va-i-v!
õre
spice she-POSS eat-IF air-CL.BIG.ROUND pain-HAB-ST-N.PST.3.AN she-OBL
13
Morse & Maxwell (1999) add ‘desiderative’ and ‘frustrative’ as morphemes that derive stative and eventive stems,
respectively. I do not include these morphemes in (27) because they are not suffixes, but full verbs, what is clear by
their syntactic characterisitics.
14
Unfortunately, due to space limitations I will be unable to show the full set of forms for the TEMP categories.
‘she always has indigestion if she eats chili’
lit. ‘if eating chili, her heart always hurts to her’
The logical relation between “eating” and “hurting” is the same for both sentences, with
and without the habitual. An explanation for this aspectual shift can be related to the fact that the
form headed by a stative verb root (20a), that lacks the habitual morpheme, has an adverbial
adjunct that causes the semantic shift from the more common interpretation in (19a) to the
marked interpretation in (20a).15
Stable states are more commonly formed by two types of constructions. Below I show
examples of the type marked by a copula, where the meaning of the forms corresponds to
Portuguese and Spanish use of the copula ser 16.
(21)
a.
i-ye
ã-i-ye
heme-bo
hia=bu
this-IN.MS eat-ST-IN.MS agouti-CL.OVAL flesh=n.3.COP .AN
‘this food is agouti meat’
b.
i-ko
b!k!-h$-ko
hí-y&kõ=me
this-FEM old-DIM-FEM MY-grand.mother=3.AN.SG.COP
‘this little old woman is my grand mother’
The second construction type of stable states is formed by an eventive verb root, -i
‘stative’ and non-past TEMP morphemes. This structure is the same used for states of ongoing
events (cf. section 4.2), although they have clear semantic differences, which justifies a separate
treatment of them. The following sentences are stable states:
(22)
a.
hí-mak! mea-rõ upa-i-bi
MY-son good-IN.C sing-ST-N.PST.3.MSC
‘My son sings well’
b.
õ pami-e
yáwa-i-biko
she kubeo-IN.MS speak-ST-N.PST.3.FEM
‘She speaks Kubeo’
c.
yãi
yawi h!óiye
hebe-i-bi
this.MSC jaguar heal-ST-IN.MS finnish-ST-N.PST.3.MSC
‘this shaman masters every healing’
The forms above are generic statements about a particular subject, which do not need to
be tied to a specific event, but are seen as a general property of the subject. This is an interesting
fact for cross-linguistic comparison, since forms in (22) are usually homophonous with habituals
or generic sentences cross-linguistically.
In Kubeo, they are different in form and meaning from habituals: semantically, the latter
must be based in an observation of constant, routine and regular event. It is a generalization over
a set of a customarily repetition of an event. Grammatically, habituals are coded by specific
morphemes and are perfective predicates (see examples further below).
Also, forms in (22) are distinct from generic sentences in Kubeo. Semantically, generics
have a meaning that implies an universal statement and are not tied to a specific event nor to
15
Shift of this kind is common with Kubeo aspectual categories. Actually, aspectual semantics is notoriously open to
semantic shifts and meaning coercions (cf. Moens & Steedman, 1988).
16
Which only codes stable states.
specific referents. It is generally used to code information about classes of beings, such as birds,
animals, human beings, spirits, etc. Generic sentences are also grammatically distinct from
habituals and stable states in Kubeo. They are formed from a stative stem and past tense TEMP
suffixes. Consider the examples below:
(23)
a.
yawi hau a-be-ame
jaguar bark do-NEG-PST.3.MSC
‘jaguars do not bark’
b.
õ-i
kuã-yo
aru õ-i
hipo-b!-re
hapu-i-ama
she-POSS bone-cl.long and she-POSS head-CL.TUBE-OBL blow-ST-PST.3.PL
na b!k!-wa
they old-PL
‘our elders/ancestors play songs with her (a deer’s) bones and skull’
c.
kaparo
ã-i-ame
yaim! he$-ye-de
aru yoka-re
wooly.monkey eat-ST-PST.3.MSC vine
fruit-IN.MS-OBL and leaf-OBL
ã-i-yame
!"
eat-ST-PST.3.MSC he
‘the wooly monkeys eat vine fruits and leaves’
Interestingly, sentences (23b) can have one interpretation that is limited to past time
reference, as with the meaning of b!k!wa referring to ‘ancestors’ instead of elders. If this is so,
then either past or present generic sentences have the same structure. This reinforces the claim
that the meaning of generic sentences in Kubeo is not limited by specific nominal or temporal
referents, but implies a universal statement over classes of beings and things.
Kubeo distinguishes two types of habituals. One is concerned with relative present time
habitual acitivities and the other with relative past habits. In fact, the latter can be analyzed
simply as a durative morpheme for past situations. Both have different morphological makeup
and will be discussed separately.
First, it must be said that habituals in Kubeo are treated as some kind of perfective
category, since stems formed by the habitual morphemes present the same aspectual semantics of
other eventive stems, as in the English sentence ‘I worked in that factory for twenty years’.
The present habitual is formed by a verb root, the habitual suffix –wa and non-past tense
suffixes. The habitual morpheme precedes every other verbal suffix, except for the negative
suffix. Consider the forms below:
(24)
a.
hi-pak!
kai=n!m!-a moa boa-wa-i-bi
MY-father every=day-PL fish kill-PSNT.HAB-ST-N.PST.3.MSC
‘My father is always/usually fishing’
b.
!h!
ba-e-de
wek!-wa hia !ra-re korika dá-wa-ma
summer be-MS-OBL tapir-PL river big-OBL middle come-PSNT.HAB-N.PST.3.PL
‘during the summer, the tapirs usually come to the Vaupes river’
There is subtle difference between the habitual activities above: while (24a) has a stative
suffix (24b) do not. This seems to be related to a contrast in the situation type description, rather
than in a different type of habitual activity. For instance, to say that ‘my father fishes every day’
is slightly different from saying ‘my father is out fishing every day’. The sentences below can
illustrate this from another perspective:
(25)
a.
- ye-de
mi-pako
yá-wa-ri
koeda-yo ?
what-OBL YOUR-mother do-PSNT.HAB-N.PST.Q awake-SIM.FEM
- waru-bo
toahi-yá-wa-biko
fish.soup-CL.WET hot-MAKE-PSNT .HAB-N.PST.3.FEM
lit. ‘What does your mother usually do when she is awake? She usually warms up the fish soup’
b.
áipe yá-wak-!-r!!
m! koeda-y!
what do-PSNT.HAB-ANT.MSC-2.Q you awake-SIM.MSC
y! kuya-wa-i-w!
I bathe-PSNT.HAB-ST-N.PST.N.3.AN
‘I am usually bathing (when I am awake)’
Since both questions do not have the stative suffix they do not frame the answer to be
related to a particular kind of habitual activity described as stative, what shows that the use of –i
‘stative’ in (25b) describes not a type of habit, but a type of situation.
The fact that the habitual suffix forms an eventive stem is attested by the
ungrammaticality of using –te ‘dynamic’ after –wa ‘habitual’ within the same stem. As it is
shown in section 4.4, -te ‘dynamic’ can only be combined to stative stems and nouns.
Habitual activities in the past are formed by the durative suffix -reha and past TEMP
suffixes. Consider the forms below:
(26)
a.
beha-reha-ima
pai-wa
go.down-PST.HAB-PST.3.PL
priest-PL
‘The priests used to go down river (a long time ago)’
b.
hi-pak!-re
káte-deha-kak!
!!re
í!-y!!-re
MY-father-OBL help-PST.HAB-PST.1.SG
pupunha get-SIM.MSC-OBL
‘I used to help my father to collect pupunha (fruit sp.)’
c.
y!hã
n!-reha-karã
no bue-kiye
hipoka
we.EXCL go-PST.HAB-PST.1PST.EXCL there study-POST.IN.MS before
‘We used to go there before classes started’
Despite the fact that all of the sentences above code some kind of habitual activty in the
past that are no longer taking place, –reha ‘past durative’ not only codes habituality, but also
other types of durative predicates. This is desirable since the past TEMP forms do not leave open
the possibility for stative stems to code durative situations, since all stative stems with past TEMP
suffixes always code generic predicates. The reason why durative situations are perfective and not
imperfective in Kubeo is a problem of temporal reference, which is discussed in section 4.6.
4.2 Progressive vs. states of ongoing events
Progressives in Kubeo are formed by an eventive verb with simultaneity nominalizers (cf.
section 3) and a copula. States of ongoing events are formed by an eventive verb root plus –i
‘stative’ and non-past TEMP morphemes. Consider first the progressive forms in (27):
(27)
a.
moá boa-y!=be
fish kill-SIM.MSC=3.AN.SG.COP
‘he is fishing’
b.
kawa !m!-i
w!-y!-re hã-y!!=mu
vulture high-LOC fly-SIM.MSC-OBL see-SIM.MSC-N.3.AN.COP.
‘I am looking at a vulture flying in the sky’
c.
hió
eda-yo=bu
garden arrive-SIM.FEM=N.3.AN.COP
‘I am arriving tired from the garden’
d.
hí-pak! buçi
nu-y!"=me
MY-father tobacco suck-SIM.MSC=N .3.AN .SG.COP
‘My father is smoking’
!et!-yo
tired-SIM.FEM
In (27a) and (27d) one finds atelic activity events with incorporated object17. In (27b) the
English translation captures the semantics of the particular Kubeo verb form, since ‘looking at’
describes the situation with stages (as Vendler 1957 use the term to defines activities).
(27c) is a very interesting example. The subject is already in the place of his/her arrival,
but there is something more in the meaning of ‘arriving’ than just getting into the designated
arrival place. The progressive with the achievement verb eda ‘to arrive’ describes the stages that
correspond to the pragmatic intuitive knowledge after ‘arriving’, such as opening the house’s
door, leaving yucca in the kitchen, greeting people, putting the shotgun on the wall, etc.
Therefore, in (27c) the Kubeo progressive differs from the common interpretation of
progressives and achievements cross-linguistically, which usually imply a preparatory stage
before the telic point. This is one point where the progressive is different from the forms of states
of ongoing events, since the latter describes events prior to its telic point (see further below).
States of ongoing situations describe particular instants of a situation before its terminus
or telic point. There is a difference in meaning depending on whether the verb root is telic (as in
(28)) or atelic (as in (29)). Consider the examples below:
(28)
(29)
17
a.
!!
moa boa-i-bi
she fish kill-ST-N.PST.3.MSC
‘he is about to catch a fish’
b.
kari-de eda-i-biko
curr.tiem-OBL arrive-ST-N.PST .3.MSC
‘she is about to arrive now’
c.
!! nomi
ea-i-bi
he woman find-ST-N.PST.3msc
‘he is persuading/attracting a woman’
a.
oko koaka-i-w!
water boil-ST-N.PST.N.3.AN
‘the water is almost boiling’
b.
y! hã%-i-w!
kawa-re
!m!-i
I look-ST-N.PST.N.3.AN vulture-OBL high-LOC
‘I see a hawk in the sky’
In (26a) and (26d) the lack of case marking and a particular pitch contour between the incorporated noun and the
verb (the same contour that occurs with compounds) support this analysis.
c.
oko
da-i-w!
water come-ST-N.PST.3.AN
‘rain is coming’
A crucial point about the sentences above is that despite the fact that translations express
some kind of progressivity or dynamicity, they are all states. One good reason for making this
claim is that progressives can be used as answers to the question ‘what is x doing?’, while states
of ongoing events cannot be possible answer to that question, what suggests that progressives are
essentially dynamic, while states of ongoing events are essentially stative. Moreover, crosslinguistically imperfective aspect in general can be used as answers to the question of ‘what is X
doing?’, such as the imperfective past tense forms in Romance languages. This makes the
morpheme –i ‘stative’ a particularly interesting case for typological and theoretical studies of
aspect.
As states, the sentences above refer to a particular instant of the situations being
described, as Vendler (1957) uses the term instant to classify states in general. In (28), all
situations are described as imminent, as a point in the time structure of the verbs prior to the
telic event, but at some point in the progress of the activity. They are all atelic, and there could be
some independent event that could cancel the completion of the situation.
Therefore, (28a) refers to a point in the fishing activity exactly when the fisherman feels
the fish in the hook or just before from pulling it from the water 18. Fishermen usually say boa-iw! (kill-ST-N.PST-N.3.AN) ‘I got it’ when a fish pulls the hook (even though the fish is still in the
water and had not been “completely caugth”). Sentence (28b) refers to a moment where there is
visible sight of a person coming to a particular place, though that person has not arrived yet.
Sentence (28c) is particularly interesting: without –i ‘stative’ it would mean that the man
has already found and persuaded a woman to be his wife, though the particular form above refers
to a moment where he is still trying to persuade or attract her.
The situations in (29) are all non-telic and refer to particular moments or a single stage in
the activities. Sentence (29a) is interesting in respect to evidentiality: if a person is present in the
place where ‘boiling’ is happening, he must use the construction in (28d). However, if the person
has moved away from the place where he had observed ‘the water boiling’, than he must use a
sentence without –i ‘stative’. The verb form in (28b) is usually used when there is a group of
people searching for something, and then the first person to catch sight of that thing would say
hã-i-w! ‘I see it’. Therefore, it refers to the very first stage of ‘seeing’ something. (28c) refers to
the moment when a person see black clouds in the sky.
It is an interesting fact that Kubeo distinguish true progressives as in (27) from states of
ongoing events, as in (28) and (29). This is a remarkable fact for theories of aspect.
4.3 Resultatives vs. non-past perfective events
Resultatives in Kubeo are of two types: non-past and past resultative. Non-past
resultatives are formed by the an eventive verb and a set of anterior nominalizers (cf. section 3)
plus a copula. Past resultative is formed by the finite suffix –kemaw!.
There are two types of non-past resultative constructions: the experience resultative,
which highlights the subject as an experiencer or agent; and the contingent resultative that
highlights more the resultant state of a prior event, rather then the referent argument as an
experience, which has a low or zero agentivity status.
The former only occurs with animate beings and always agrees with the gender and
person of the subject. I will first discuss the experience resultative. Consider the examples below:
18
The particular intepretations of this and other sentences in (48) were extensively discussed with Kubeo
consultants during fielwork.
(30)
a.
yawimi a!p!-k!
ea-k!=be
dog
alive-ANT .MSC get-ANT.MSC=3.AN.SG.COP
‘the dog arrived in the house alive (after being shot)’
b.
oro ea-k!=bu
gold get-ANT.MSC=N.3.AN.COP
‘he has found gold’
c.
mahe
y&k#yo
yai-k!=be
our.INCL grand.father die-ANT.MSC=3.AN.SG.COP
‘Our grand father died yesterday’
k!rami-i
house-LOC
enoa-re
yesterday-OBL
While (30a) and (30b) have agent subject, (30c) has a patient of a state. Nevertheless the
form above highlights it as an experiencer.
Interestingly, the structure of the experiencer resultative is homologous to the Present
Perfect forms cross-linguistically, where the main verb is nominalized with a past participle form,
and the auxiliary has present or non-past tense.
On the other hand, the contingent resultative does not necessarily agrees with the gender
and person of the referent argument. It is marked an eventive verb root, followed by oblique case
marking (the inanimate marker of the eventive verbs paradigm of anterior nominalizer in section
3.1) and an invariant clitic copula, homophonous with the clitic copula used for non-3rdperson
animate subjects.
This type of resultative is more commonly found with inanimate argument referents. This
also corresponds to the fact that inanimates are low-class arguments on the agentivity scale, so
there are possibly semantic constraints on using inanimate referents as subjects of the experiencer
resultative. Consider the sentences below:
(31)
a.
!#yei ñemi-ne=bu
grape black-OBL =3.AN.SG.COP
‘the grape has ripened’
b.
õpõ-y!! po-be-Ø-w!
ame-te-de=bu
gun-cl.? explode-neg-N.PST-N.3.AN bad-DYN-OBL =N.3.AN.COP
‘the gun does not shoot, it has damaged’
c.
ph!
wea-re ãr!-me-te-de=bu
corn-OBL think-NEG-DYN-OBL =N.3.AN.COP
‘Oh, I have forgotten the corn!’
INTRJ
d.
ñ!hã
kã-r"
n!-re=bu
paí-de
hápi-kari
we.EXCL sleep-CONV go-OBL = 3.AN.SG.COP priest-OBL listen-ANT.NMZ
‘we have slept after listening to the priest’
The sentence in (31a) is a pattern used commonly for ripening, blooming and other
similar natural phenomena. The sentence in (31c) is interesting for giving descriptions of two
types of situations: in the first sentence there is a present momentary statement about a gun,
which is followed by conclusive statement about its state. Sentence (31c) and (31d) have animate
first person referent arguments, though there is no agreement between them and the verbs. In all
cases, referent arguments have no control of the events.
The past resultative has strong similarities with the contingent non-past resultative. Both
have no distinct agreement pattern for animate and inanimate referents.
This is particularly important because Kubeo (and Tukanoan languages generally) have
systematic nominative-accusative agreement marking at NP and VP levels.
Consider the following past resultative sentences below:
(32)
a.
yu-i
k!-há-k!
aru-ta
here-LOC exist-IMPR-MSC and-FOC
‘stay here!’ he said, so we stayed’
b.
na da-kemaw!!
moã-w!
they come-PST.RSLT fish-an.pl
‘they came as fish’
ñ!ha
k!#-te-kémaw!!
we.EXCL exist-DYN-PST.RSLT
Comrie (1976:20) mentions that while a resultative interpretation can be inferred from the
perfective aspect, this is not required to occur. In Kubeo, this also true, but the resultative
constructions contrast semantically with perfective forms with non-past tense forms. This contrast
can be illustrated by the following sentences:
(31)
a.
eda-rã-ma
arrive-ANT .PL.AN-3.PL.COP
‘they have arrived’
b.
eda-ma
arrive-N.PST.3.PL.AN
‘they have arrived’
Although both are near synonyms, Kubeo speakers assert they are different. The way
native speakers express the contrast is the following: (31b) can be uttered when one sees the
arriving people in the river port, while (31a) is uttered when some one sees the people that he/she
was expecting to arrive in a house, in the village center, or in some place and time not immediate
to where and when the arrival took place. This sense of non-immediate experience is because the
resultative forms refer to the resultant state after the situation took place and not necessarily with
the situation itself.19
4.4 The morpheme -te ‘dynanmic’
It could be wrongly supposed that the morpheme –te ‘dynamic’ is a marker of perfective
aspect20. This ‘suffix’ actually functions as an independent verb too, with roughly the same
meaning as the English verb ‘to do’ with a weak transitive meaning, such as in the sentence “I am
going to do the roller-coasters and you are going to do shopping”. In Kubeo, there is a similar use
of te as an independent verb, illustrated in (32):
(32)
19
a.
hio-i
te-w!
Morse & Maxwell (1999) have analyzed resultatives as Assumed Evidential. This is because in resultatives there is a
sense of temporal and spatial distance between the situation and its resultant state. However, they are by no means an
evidential category. There are two main reasons against an evidential analysis: there resultatives about the speaker, who
has both uttered and experienced the situation. Also both resultatives and non-past perfective forms are statement based
on first hand experience of the speaker; they just describe the situations in two different ways.
20
Likewise, it wrongly be thought that –i ‘stative’ is a marker of imperfective aspect. However, we have already seen
that –i ‘stative’ function in other domains of Kubeo grammar, and not exclusively on aspect, such as in syntax and
nominalizations.
garden-LOC
do-N.PST.N.3.AN
‘I have worked in the garden’
The fact that te ‘do’ can function independently and also as a bound morpheme is a
peculiarity of Kubeo agglutinating grammar and diachronic processes. Morphophonoligcally the
two functions are distinct: te as a verb is stressless, while –te as suffix can be stressed depending
on the metrical structure of the word it belongs to.
As a suffix, –te ‘dynamic’ can only get combined with stative verb roots, stative stems
and nouns. The following examples illustrates this latter use:
(33)
a.
p!ka-rã
põé-te-ma
two-AN.PL
person-DYN-N.PST .3.PL
‘two babies have been born’
b.
b!k!-k!
wei-te-bi
old-ANT.MSC black.paint-DYN-N.PST.3.MSC
‘the old man has gotten painted black’
The dynamic suffix interacts with the meaning of the noun root in creating an inchoative
predicate. In contrast, in the causative predicate below, what was once the subject of the
inchoative sentence in (33a), becomes the object:
(34)
a.
p!ka-rã-re
põé-te-wa-biko
two-PL-OBL
person-DYN-CAUS-N.PST.3.FEM
‘she gave birth to two babies’
Therefore, while the semantics of –te indicates dynamism, it does not indicate anything
related to agentivity. From examples (34) it is also excluded the possibility of analyzing –te
‘dynamic’ as an exclusive perfective marker.
4.5 Stative stems, negation and inchoative predicates
A stem that is headed by the negative suffix -be has the same semantic an grammatical
pattern of other stative stems. For instance, an eventive verb that normally yields a perfective
interpretation, as in (35a), with negation it yields an imperfective interpretation, as in (35b). Also,
in (35c) one finds a negated predicate with a stable state meaning. In (36) negation is the head of
a stem in a predicate with generic meaning:
(35)
(36)
a.
kari
hawe-hi-na-re
oka-w!
curr.time already-DIM-?-OBL rain-N.PST .N.3.AN
‘it has rained this morning’
b.
kari-de
oka-be-w!
curr.time-ACC rain-NEG-N.PST.N.3.AN
‘it is not raining now’
c.
koh!-be-w!
di-ye-de
like-NEG-N.PST.N.3.AN ANPH-IN.MS-OBL
‘I don’t like that’
!-be-ni
k!-ri
hãrãw!-re yáwa-be-ame
pamu-r!
want-NEG-CONV exist-NMZ day-OBL speak-NEG-PST.3.MSC armadillo-CL.ROUND
m!a-i-ta
yáwa-i-yame
spring-LOC-FOC speak-ST-PST.3.MSC
‘although he does not sing every time of the year, he sings during the armadillo’s spring season’
It is possible to make eventive stems after a negation suffix. We can compare (37) below,
marked by negation followed by –te ‘dynamic’, with sentence (35b) above:
(37)
kari
hawe-hi-na-re
oka-be-te-w!
current.time already-DIM-?-OBL rain-NEG-DYN-N.PST.N.3.AN
‘it has not rained this morning’
It is also possible to add –i ‘stative’ and then the predicate would be imperfective:
(38)
da-be-te-i-ma
come-NEG-DYN-ST-N.PST.3.pl
‘they have not been coming’
As one can see, the predicate aspect is compositional: –te ‘dynamic’ indicates a relevant
left boundary of the event in the past, while –i ‘stative’ codes durativity. This type of
compositional meaning support Smith 1997 approach, and suggest that while the right most head
of a stem has wider scope over the preceding head, it does not cancel the aspectual properties of
the previous head, but the predicate shows a composition of aspectual meaning from the two
heads. This can be better illustrated with the inchoative predicates below:
(39)
a.
kari-de
toahi-te-i-w!
curr.time-OBL hot-DYN-ST-N.PST.N.3.AN
‘it is getting hot’
b.
ihi-m!
má-k!
mahi-ye
mea-te-i-w!
y!
pain-PASS.MSC be-ant.msc little-in.MS good-dyn-st-N.PST.N.3.AN I
‘I was sick, but I am getting a little better’
Without the stative suffix, the sentences above would imply a perfective meaning, such
as ‘now it got hot’, or ‘I got healed’, respectively. However, with the stative suffix the predicates
show a composition of INCHOATIVE + IMPERFECTIVE, i.e not a full culmination of the inchoative
predicate, but its starting point and progress towards the culmination point.
Inchoatives always occur in the composition of stative root or noun and –te ‘dynamic’.
However, it is possible to have also a closed state interpretation as illustrated in (10) above.
Below, I give three examples with the same verb form that code a Perfect predicate (40a), a
closed state (40b) and an inchoative predicate (40c).
(40)
a.
y! ihi-ye-te-w!
yóbek!ri hãrãw!-wa aru ihi-Ø-w!
I pain-IN.MS-DYN-N.PST.N.3.AN three
day-PL
and pain-N.PST-N.3.AN
y!-re kari-de
I-OBL current.time-OBL
‘I have been sick for three days and I am (still) sick’
b.
yóbek!rã hãrãw!-wa
three
day-PL
ihi-ye-te-w!
enoamare-de
pain-IN.MS -DYN-N.PST.N.3.AN yesterday-OBL
h!o-Ø-w!
heal-N.PST-N.3.AN
‘I was sick for three days. Yesterday I got better’
c.
ihi-ye-te-Ø-w!
enoamare-de oko ihiya-ke-de
#k#-y!
pain-IN.MS-DYN-N.PST-N.3.AN yesterday-OBL water Vaupes-ORG.MS-OBL drink-SIM.MSC
‘I got sick yesterday drinking water from the Vaupes river’
4.6 Temporal reference
There are no clear temporal boundaries between past and non-past TEMP morphemes
when the situation time is in the recent past time reference. The following sentences have the
same temporal reference, nevertheless they have different TEMP forms:
(41)
a.
õ-i
n!-ki
hipoka moa h!y!-wa-r$
te-w!
she-POSS go-POST before fish smoked-CAUS-CONV do-N.PST.N.3.AN
‘before she went, I smoked fish (for her travel)’
b.
õ-i
n!-ki
hipoka !ra-rõ y!-re h!t!ra
karo-r$
hí-ako
she-POSS go-POST before big-IN.C I-OBL flour
toast-CONV give-PST.3.FEM
‘before she went, she gave me a lot of toasted flour’
It is hard for now to give a precise definition of temporal reference between the two
TEMP categories. It could be that in non-past TEMP there is a feeling of present relevance, while
there is not for the past TEMP. In narratives of historical or mythological times, it is more
common to find past TEMP markers instead of non-past suffixes. Also, in generic sentences, the
past TEMP forms can also appear in atemporal predicates.
Moreover, non-past TEMP forms have a vague temporal reference definition, and
depending on the context they can appear in present, past or future time reference. Consider the
sentence below, where an eventive stem marked by non-past tense may have an interpretation
similar to a present Perfect meaning21:
(42)
hí-mak! !re moa boa-bi
MY-son very fish kill-N.PST.3.MSC
‘my son has caught a lot of fish’
An eventive verb root with past TEMP forms always code a situation in the past.
However, with non-past TEMP forms it can also code situations in the future:
(43)
a.
b.
21
hawena !!-i
kerari n!-i-ye-de
#k#-i-no
má-te-w!
tomorrow he-POSS Querari go-ST-IN.MS-OBL drink-ST-IN.C be-DYN-N.PST.N.3.AN
‘Tomorrow when he goes to Querari there will be a party over there’
lit. ‘Tomorrow at his going to Querari there will have been a party (over there)’
hawena y!hã
waru-e-de
mia-wã
eda-ma
tomorrow we-EXCL fish.soup-IN.MS-OBL mosquito-PL arrive-N.PST.3.PL
‘tomorrow during our (morning) fish soup, mosquitos will arrive’
One crucial difference, perhaps, between the Kubeo form above and those reported for Niger-Congo
languages is that in Kubeo such a statement implies that the speaker was a witness of the situation. This is
due to the implicit evidentiality in non-past forms.
The only time reference constraint of perfective forms is present time, which follows
from pragmatic constraints as it was discussed in section 2 above.
Forms with copulas can also have future time reference, as in (44a) and (44b):
(44)
a.
n!-ñ!=mu
kopai da-be-k!=bu
go-SIM.MSC-N.3.COP return come-NEG-ANT.MSC-N.3.AN.COP
‘I am going and I am not coming back’
b.
y!"h&
kã-r$
koeda-i-ye-de
oyo-wa eta-i-na-ma
our.EXCL sleep-CONV wake-ST-IN.MS-OBL bat-PL leave-ST-ANT.AN.PL-3.PL
‘when we wake up the bats will be leaving’
This shows that these forms resemble imperfectives cross-linguistically in being able to
refer to a future situation. The similarity of the resultative form in (44b) with the Perfect is clear
where the situation of ‘bats leaving’ will have started before the event of ‘our waking up’, and it
will still continue after that.
Imperfective predicates in Kubeo can only have present and future meaning, while
perfective predicates can have past, Perfect and future meaning, i.e. anything but not present. It is
remarkable that forms in the past must all have a perfective meaning. Durativity in the past
TEMP is marked by an eventive stem.
The interaction of the aspect of the stems and the TEMP categories suggest that non-past
is a finite set of person, evidentiality and mood categories that has no constraint in time reference.
Therefore, time is underspecified, which suggest its negative classification as non-past TEMP22.
Aspect, adverbials and the context are the main factors responsible for time reference in non-past
TEMP.
Time seems to be specified for past TEMP, but there some problems. First, stative stems
with past TEMP code generic predicates. Second past TEMP seems to code not only temporal
distance from the speech time, but also spatial distance from the speech place. I have observed
the use of past tense forms for situations that are true for the moment of speech, but the person
that is being spoken about is far away in another city, village, etc. Consider the sentence below,
which was used by a Kubeo speaker to correct a statement that I did using the non-past tense:
(45)
hí-ma-re
pako mika-te-ako
my-children-OBL mother be.pregnant-DYN-PST.3.FEM
‘my wife is pregnant’
Other morphological evidence suggests that the forms that mark past TEMP have close
morphological relations to noun and adjective forms23. This morphological facts and the
correlation of spatial and temporal distance in the meaning of past TEMP forms might explain the
fact that generic sentences and past situations are coded by the same TEMP paradigm. More
investigation on these issues is extremely necessary.
5. Conclusion
In this paper I presented a correlation in Kubeo grammar of the lexical aspect of verb
stems and the viewpoint aspect of predicates. It was shown that Kubeo does not have
22
Future constructions using non-finite forms will not be discussed in this paper.
For instance, the morphme –ka appears in several forms of the past TEMP and also in constructions that express the
place of origin of something. Also, person endings in past TEMP are also found in non-finite forms and noun and
adjective inflection.
23
grammaticalized aspect, so notions such as perfective and imperfective are inferred from the
meaning of predicates. The sole grammatical fact responsible for aspect in Kubeo is the stem
level, which is formed by the composition of the lexical aspectual properties of verb roots and
non-category changing derivational morphemes in the stem.
A correlation of stativity with imperfectivity, and eventivity with perfectivity (cf. (12))
was proposed semantically and demonstrated grammatically in Kubeo.
Kubeo predicates have a wide range of aspectual categories and situation descriptions.
The correlation of root semantics, morphological heads (–i ‘stative’ and -te ‘dynamic’), modality,
valence and copular constructions in a complex aspectual system is an interesting convergence of
grammatical and semantic factors in the study of aspect in general. The grammatical function of –
i ‘stative’ (and also –te ‘dynamic’) and the types of predicates it code are unique typologically
and relevant for a theory of aspectuality.
Tense is a category in the language that does not seem to be so grammaticalized as lexical
aspect, although there is a binary past/non-past tense opposition. The language is an interesting
case for the study of the diachronic evolution of tense systems. More broadly, temporal reference
is achieved by several grammatical and lexical devices, especially lexical aspect.
Bibliography
Bach, E. 1986. The Algebra of Events, Linguistics & Philosophy 9, 5-16.
Binnick Robert I. 1991. Time and the Verb: A Guide to Tense and Aspect. Oxford University
Press.
Carlson, Gregory N. 1977. A unified analysis of the English bare plurals. Linguistics &
Philosphy 1.3:413-456.
Comrie, Bernard. 1976. Aspect. Cambridge: Cambridge University Press.
______. 1985. Tense. Cambridge: Cambridge University Press.
Cook, Dorothy and Linda L. Criswell 1993. El idioma koreguaje (Tucano occidental). Bogotá:
ILV.
Csirmaz, Aniko. 2005. Semantics and phonology in syntax. PhD dissertation, Massachusetts
Institute of Technology. Distributed by MITWPL.
Dahl, Osten. 1985.Tense and aspect systems. New York: Basil Blackwell.
Emenanjo, Emmanuel !Nolue. 1991. The tense system of Igbo. AAP 27.
Declerck, Renaat. 1979. Aspect and the bounded/unbounded (telic/atelic) distinction. Linguistics
17, 761-794.
Declerck, Renaat. 1979. Boundedness and the Structure of Situations, Leuvense Bijdragen 78,
275-308.
Depraetere, Ilse. 1995. On the necessity of distinguishing between (un)boundedness and
(a)telicity. Linguistics and Philosophy 18, 1 -19.
Filip, Hana & Carlson, G. N. 1997. Sui Generis Genericity. Penn Working Papers in Linguistics
4. Philadelphia: The University of Pennsylvania. 91-11.
Fleischman, Suzanne. 1995. Imperfectives and irrealis. In Modality in Grammar and Discourse,
ed. by Joan Bybee and Suzanne Fleischman, 519–551. Amsterdam and Philadelphia:
John Benjamins.
Gawronska, Barbara. 1993. An MT Oriented Model of Aspect and Article Semantics. Lund
University Press.
Givon, T. 1984. Syntax: a functional Typological Introduction. Volume I. John Benjamins.
Hatav, G. 1989. Aspect, Aktionsarten, and the Time line. Linguistics 27. 487-516.
Hollinger, Cari, Linda Criswell and Nancy L. Morse. 2000. El cubeo. In: María Stella González
de Pérez and María Luisa Rodríguez de Montes eds. Lenguas indígenas de Colombia: una
visión descriptive. 357-72. Bogotá: Instituto Caro y Cuervo.
Johnson, Orville and Stephen Levinsohn 1990. Gramática secoya. Cuadernos Etnolinguísticos 11.
Quito:ILV.
Kiparsky P. 1998. Partitive case and aspect. in M Butt & W Geuder eds. The Projection of
Arguments: Lexical and Compositional Factors CSLI Stanford: 265-307.
Klein, Wolfgang. 1994. Time in Language. London: Routledge.
Mason, J. A.: 1950, The Languages of South American Indians,Bur. Amer. Ethnol. Bull. 143 v. 6,
157–317 and map.
Moens, M. and M. Steedman. 1988. Temporal ontology and temporal reference. Computational
Linguistics, 14:15-28.
Morse, Nancy & Michael B. Maxwell. 1999a. Cubeo Grammar. (Studies in the Languages of
Colombia, 5; SIL Publications in Linguistics, 130.) Arlington: SIL/U of Texas Arlington
Press.
Mourelatos, Alexander P. D. 1978. Events, Processes, and States. Linguistics & Philosphy 2. 415434.
Mufwene, Salikoko. 1984. Stativity and the progressive. IULC.
Rothstein, Susan, 2004: Structuring Events: A Study in the Semantics of Lexical Aspect.
Blackwell: Oxford.
Smith, C. S. 1997. The Parameter of Aspect. Kluwer, Dordrecht.
Velie, Daniel. 1975. Bosquejo de la fonología y gramática del idioma orejón (coto). Datos EtnoLingüísticos,
10.
Lima:
Instituto
Lingüístico
de
Verano.
32
p.
http://www.sil.org/americas/peru/show_work.asp?id=12575
Vendler, Zeno. 1957. Verbs and times. Philosophical Review 56:143–160.
Welmers, William Everett. 1977. African Language Structures. Berkeley: University of
California.
6. Final Note
Center for American Indian Languages
University of Utah
255S.Central Campus Dr.
Linguistics,Rm.2300
Salt Lake City, Ut 84112-0492